or film printing with (2) data reduction and analysis consisting of 
image enhancement, spectral analysis, signature correlations, and 
recognition computations. Care in understanding the scope of data 
processing is required because the term can mean only (1) above, 
thereby leaving out data reduction, analysis and recognition (or classi- 
fication) which are critical to the user as information extraction 
processes. 
My assessment of the status of processing techniques is as 
follows: 
l. A variety of techniques have been demonstrated to be 
feasible in many applications under limited conditions 
appropriate to showing feasibility but not generally 
appropriate to prototype operational conditions. 
a. Constrained data collection to minimize effects 
of resolution, atmosphere, and changing illumina- 
tion. 
Little or no time constraint for processing has yet 
been imposed. 
c. Maximum ground observation has been used. 
These limitations are being lessened to the point where 
operational-prototype information systems are feasible 
in some applications. 
3. Further processing technique development is 
necessary. 
The transferral of present techniques to prototype environments is 
needed and requires effort to solve problems which arise that are not 
otherwise evident. The Corn Blight Watch experiment pressured condi- 
tions (lb) and (lc) above by imposing a time constraint of less than 
one day per 16 square kilometer segment processed and reported and 
reducing the amount of ground data collected. Aircraft data collection 
costs can be reduced by relaxing the constraints on sun angles and cloud 
cover percentage presently imposed by processing limitations. Further 
automatic processing technique development is clearly required. 
A discussion will be given of some of the more powerful and 
sophisticated processing techniques with the simpler techniques like 
level slicing and contouring not described despite their usefulness. 
3.2.1. Signal Conditioning (Preprocessing) Statistical design 
techniques are employed to provide useful features. In applying these 
statistical algorithms it is necessary to condition the inputs to the 
recognition device (and also the data base used to design the recogni- 
tion device). A properly conditioned data base will simplify the design 
and operation of the system. The data conditioning is performed by 
a preprocessor whose role as a component of the recognition system is 
  
29